Research Assistant
Accelerated large language model inference in llama.cpp using 16-bit activation.Pruned stable diffusion model to reduce the total flops of the neural network.
Please complete the CAPTCHA to continue
@fudan.edu.cn
✓
LinkedIn matched
A concise factual answer block for searchers comparing this professional profile.
Martin Wang is listed as Actively looking for Co-op and Internship opportunities | Google Summer of Camp 2024, OpenVINO | AI Framework Engineer @ Intel | Senior Software Engineer @ AMD | AI Model Compression, AI application algorithm development based in Greater Boston, United States. AeroLeads shows a work email signal at fudan.edu.cn and a matched LinkedIn profile for Martin Wang.
Martin Wang previously worked as Research Assistant at Northeastern University and Algorithm Engineer at Frontis Technology. Martin Wang holds Informatics, 4.0/4.0 from Northeastern University.
This section adds company-level context without repeating Martin Wang's masked contact details.
AeroLeads found 1 current-domain work email signal for Martin Wang. Compare company email patterns before reaching out.
I have an STEM major and also industry working experience in tech companies. I really want to use my technical skills to build useful software application to make people lives better and boost the efficiency of the economy. Now I'm doing project implementing AI model as another backend to web development, because intersection of software engineering and the algorithms is the most interesting.I could provide the following skills:🔧Model Compression and Deployment: As an AI Framework Engineer at Intel Corporation, I enabled and optimized various natural language processing models for the inference engine, applying techniques such as adaptive sequence length, 8-bit quantization, and kernel fusion. I also defined and improved the performance of sparse convolution kernel in the AI compiler TVM, using the ansor searching algorithm.🚀AI tooling development: accelerated the Stable Diffusion model with controlnet module using TensorRT , and added this function to a plugin for the stable-diffusion-webui. Implemented a full-stack project chatbot for users to interact with multiple models in a same api and UI.I'm excited to continue my journey in AI application engineering. I am actively seeking an internship opportunity where I can contribute my expertise and continue to learn from the best. Let's connect and explore opportunities to collaborate and make a positive impact in the world of AI and technology!
Listed skills include Programming, Data Analysis, Matlab, Linux, and 3 others.
A career timeline built from the work history available for this profile.
Boston, Massachusetts, United States
Accelerated large language model inference in llama.cpp using 16-bit activation.Pruned stable diffusion model to reduce the total flops of the neural network.
Beijing, China
Trained the Stable Diffusion controlnet on the dataset scrapped from Internet, supporting depth, canny and gesture finer control of the diffusion model for the company’s flagship human model generation product.Investigated quantization, pruning, caching, distributed inference machine learning algorithms for Stable Diffusion models. Compiled a Stable Diffusion ControlNet engine in Nvidia TensorRT for Hugging face Diffusers and Stable Diffusion WebUI.
Shanghai, China
Compressed and quantized the neural networks by applying low-bit quantization and model distillation. Saved memory consumption and boosted perf by 20% for the transformer-like model.Developed new features in inference engine like adaptive sequence length of MiniLM Bert model in IMex(Intel Extension for Transformer) by adding compiler pass to do kernel fusion and writing high-performance kernels. Optimized the sparse matrix multiplication and convolution kernel utilizing the auto-tuning searching algorithm in AI compiler TVM. Deep-dived in flash-attention code in OpenAI Triton language for future development.
Familiar with AMD deep learning ROCM software stack and deployed trained model on AMD GPU.Maintained the GPU kernel space driver stack by debugging issues on the new GPU chip for features like command submission and memory management, and supported AMD ROCM heterogeneous architecture.
Software Engineer: Developed diagnostic tools for storage devices like ethernet card and NvMe drives for storage platforms.
Shanghai City, China
Cross function communication within and between vendors to troubleshot issues of hardware platforms.
Activities and Societies: Mission Hill volunteering and community walks, morning coffee chat on Sunday.
Quick answers generated from the profile data available on this page.
Martin Wang is listed as Actively looking for Co-op and Internship opportunities | Google Summer of Camp 2024, OpenVINO | AI Framework Engineer @ Intel | Senior Software Engineer @ AMD | AI Model Compression, AI application algorithm development.
AeroLeads has found 1 work email signal at @fudan.edu.cn for Martin Wang.
Martin Wang is based in Greater Boston, United States.
Martin Wang has worked for Northeastern University, Frontis Technology, Intel Corporation, Amd, and Dell Emc.
You can use AeroLeads to view verified contact signals for Martin Wang, including work email, phone, and LinkedIn data when available.
Martin Wang holds Informatics, 4.0/4.0 from Northeastern University.
Martin Wang is listed with skills including Programming, Data Analysis, Matlab, Linux, Mysql, Shell Scripting, and Latex.
Search by job title, company, industry, location, and seniority. Export verified B2B contact data when you need it.
Start free trial Search contactsCheck these profiles if this is not the Martin Wang you were looking for.
View similar profiles